"Serving ML Models at Scale at Allegro" - Agnieszka Rybak (MOPS Session 1)
MLOps Oriented Polish Session aims to delve into the engineering aspects of ML-powered software.
At Allegro, we face the challenge of deploying and monitoring machine learning (ML) models built by research engineers and data scientists without extensive application deployment experience. To address this, we leverage the existing DevOps expertise and tools and extend them to support not only microservices but also ML models.
In this presentation, we introduced Allegro's ML platform, which enables multiple teams to deploy custom models with low latency, robustness, scalability, and observability. We explore key components, including a custom Feature Store, Dataset Model Repository, and Online Prediction system.
For more information, follow us on LinkedIn: https://www.linkedin.com/company/mops-community/
And on MeetUp: https://www.meetup.com/pl-PL/mops-community/
Видео "Serving ML Models at Scale at Allegro" - Agnieszka Rybak (MOPS Session 1) канала MOPS Community
At Allegro, we face the challenge of deploying and monitoring machine learning (ML) models built by research engineers and data scientists without extensive application deployment experience. To address this, we leverage the existing DevOps expertise and tools and extend them to support not only microservices but also ML models.
In this presentation, we introduced Allegro's ML platform, which enables multiple teams to deploy custom models with low latency, robustness, scalability, and observability. We explore key components, including a custom Feature Store, Dataset Model Repository, and Online Prediction system.
For more information, follow us on LinkedIn: https://www.linkedin.com/company/mops-community/
And on MeetUp: https://www.meetup.com/pl-PL/mops-community/
Видео "Serving ML Models at Scale at Allegro" - Agnieszka Rybak (MOPS Session 1) канала MOPS Community
Комментарии отсутствуют
Информация о видео
11 июля 2023 г. 17:51:07
00:30:18
Другие видео канала
"Why ReAct Fails for Smaller Language Models (and How to Fix It)" - Bartosz Szaniecki (MOPS 7)
"Building machine learning platforms for a living - lessons learned" - Mateusz Kwaśniak (MOPS 2)
“Building ML Platform from Components: neptune.ai" - Aleksander Wojnarowicz (MOPS Session 1)
"Model serving 101 with FastAPI & LitServe" - Marcin Zabłocki (MOPS 6)
“MLflow iceberg: from basics to hidden depths” - Marcin Zabłocki (MOPS 3)
“Small models, big impact” - Prince Canuma (MOPS 5)
"A journey towards MLOps -from greenfield to a quite robust platform" -Wojciech Mikołajczyk (MOPS 4)
"Building production-ready LLM software" - Łukasz Myśliński (MOPS 4)
“Scalable, secure, and sustainable NatWest Group’s MLOps platform” - Marek Brynda (MOPS 3)
"Cloud-agnostic MLOps Platform with Kedro, MLflow and Terraform" - Marek Wiewiórka (MOPS 2)
"Architecting ML for Huge Impact and Scale" - Szymon Jacoń (MOPS 2)
“Building a vector search engine” - Maciej Mościcki (MOPS 3)
"Beyond the Hype: Leveraging LLMs models in creating MLOps solutions" - Piotr Kałużny (MOPS 6)
"A bag of tricks to speed up ML models inference in production" - Michał Mikołajczak (MOPS 4)
"Lego, Latency, and Life: Building MLOps on a Budget" - Agnieszka Rybak (MOPS 7)
"MLOps - A Holistic Approach" - Dariusz Kłeczek (MOPS Session 1)
“LLM Inference: Self hosting” - Vladimir Alekseichenko (MOPS 5)
"Distributed deployment of LLMs on a Datacenter Scale" - Piotr Tarasiewicz (MOPS 7)
“NVIDIA NIM: Deploying Generative AI at the Speed of Light” - Mateusz Szczesny (MOPS 5)
"Ways to Scale Up ML Experiments on AWS" - Julia Będziechowska (MOPS 6)
